Search results

Are you currently behind any proxy or VPN, such as Cloudflare? This information could help narrow down the issue. Also, could you please confirm that the server is using a unique hostname? By that I mean the hostname should be a dedicated subdomain that is not used by any of the hosted domains...

Just an idea, Have you tried to access the panel via the server IP address? and change the full hostname to a different subdomain?

Hi @Sebahat.hadzhi , would you be able to provide a patch for this issue before the Plesk Obsidian 18.0.75. update? Thank you.

What is the ETA on the fix?

Sorry for the delayed reply! Thank you very much, @Sebahat.hadzhi and @Kaspar . I really wish there were an option to permanently disable /usr/local/psa/admin/bin/statistics_collector, as it’s a long-running and unnecessary process for us. I’ve already done this, but the process is still...

Hi, I know this topic has been discussed a few times before, but is there any way to completely disable the /usr/local/psa/admin/bin/statistics_collector process? I’m running a very large website (around 600 GB) with high traffic, and I’ve noticed that this process runs daily from about 4 AM...

How about if you backup locally? does it work? Also how about the memory usage during the backup?

Also try - "Dedicated FPM application served by nginx" for your websites.

I don’t think the issue you’re seeing is directly caused by MariaDB. Given the FPM memory usage and the CPU spikes, it looks more like the server is simply running tight on resources at times. As I mentioned earlier: Fast spike + slow reclaim ⇒ the OOM killer targets the largest process...

I think .. then simply your server requires more RAM.

Ok. If I understand correctly, you are talking about buffer/cache. Not the physical memory usage by MariaDB. Page cache is global and reclaimable; MySQL can’t “flush” or “release” it— the kernel does that on demand. Fast spike + slow reclaim => OOM kills the biggest target (often mysqld). Try...

Do you mean that the MySQL memory usage itself is actually growing over 30 GB or so? ~15 GB under load is normal.

Please post your my.cnf

Bit confused. Is it 100% or 27GB of the 40GB?

It sounds like you’re describing a situation where, during the backup process, memory usage (specifically the buff/cache value) increases, and after the backup completes that memory isn’t released back to the “free” pool. If that’s the case, this is actually the expected behavior of Linux. Have...

Have you found a solution for this issue? If not, please post your MariaBD current configuration (my.cnf)

Hi @trialotto, I might be wrong, but as far as I know, Spamhaus.org does not allow direct use of their DNSBL now. We need to set up a DQS key and then use it. In my experience, using zen.spamhaus.org directly has resulted in many false positives. Thank you.

Sorry! I understand what is happening now. systemctl restart plesk-php83-fpm_* Only restart active plesk-php83-fpm_* Thank you for your help!

@Raul A Thank you! If I stop a “Dedicated PHP-FPM” via the Plesk Panel and run: systemctl restart plesk-php83-fpm_* it doesn’t bring the Dedicated PHP-FPM service back up. However, if I run my script or specify the full path, it starts running again. Example: systemctl restart...

Thank you! great idea!